Annals of Biomedical Engineering

, Volume 36, Issue 7, pp 1305–1313

Partial Least Squares Regression: A Valuable Method for Modeling Molecular Behavior in Hemodialysis

  • E.A. Fernández
  • R. Valtuille
  • P. Willshaw
  • M. Balzarini
Article

Abstract

The aim of this work was to use the Partial Least Squares Regression (PLS) technique to fit simple models for the interpretation of an underlying complex process. In this study, the technique was used to build a statistical model for molecular kinetic data obtained from hemodialyzed patients. By using PLS we derived statistical linear models for the prediction of the equilibrated urea concentration which would be reached 30–60 min after the end of the dialysis session. Models with an average relative prediction error (RPE) of less than 0.05% were achieved. The model predictive accuracy was evaluated in a cross-center study yielding an RPE < 3%. The chosen model was robust to variations such as sampling extraction time demonstrating a high capacity for modeling kinetics. It also was found to be useful for bedside monitoring. Finally, the PLS technique allowed identification of the most important co-variables in the model and of those patients with outlier patterns in their molecular dynamics.

Keywords

Kinetic modeling Equilibrated urea Hemodialysis adequacy Statistical modeling 

References

  1. 1.
    Abdi, H. Partial Least Squares (PLS) regression. In: Sage Encyclopedia of Social Science Research Methods, edited by M. Lewis-Beck, A. Bryman, and T. Futting. Thousand Oaks, CA: Sage Publications Inc., 2003Google Scholar
  2. 2.
    Canaud B., Bosc J. Y., Cabrol L. et al. Urea as a marker of adequacy in hemodialysis: lesson from in vivo urea dynamics monitoring. Kidney Int. 76(S): S28–40, 2000CrossRefGoogle Scholar
  3. 3.
    Cheng Y. L., Chol K. S., Chau K. F., Li S. A., Yung A. U., Yu A. W., Wong K. K. Urea reduction ratio that considers effects of ultrafiltration and intradialytic urea generation. Am. J. Kidney Dis. 37: 544–549, 2001PubMedCrossRefGoogle Scholar
  4. 4.
    Chong I. G., Jun C. H. Performance of some variable selection methods when multicollinearity is present. Chemomet. Int. Lab. Syst. 78: 103–112, 2005CrossRefGoogle Scholar
  5. 5.
    Daugirdas J. Simplified equations for monitoring kt/v, pcrn, ekt/v and epcrn. Adv. Ren. Replace. Ther. 2(4): 295–304, 1995PubMedGoogle Scholar
  6. 6.
    Daugirdas J. T., Depner T. A., Gotch F. A., Green T. et al. HEMO study group 1997, Comparison of methods to predict equilibrated Kt/V in the HEMO study. Kidney Int. 52: 1395–1405, 1997PubMedCrossRefGoogle Scholar
  7. 7.
    Depner T. A. History of dialysis quantification. Semin. Dial. 12(S1):14–19, 1999Google Scholar
  8. 8.
    Draper, N. R., and H. Smith. Applied Regression Analysis, 3rd edn. USA: John Wiley, 1998Google Scholar
  9. 9.
    Eknoyan G., Beck G. J., Cheung A. K. et al. Effect of dialysis dose and membrane flux in maintenance hemodialysis. N. Engl. J. Med. 347: 2010–2019, 2002PubMedCrossRefGoogle Scholar
  10. 10.
    Fernández, E. Data mining and neural networks for dialysis kinetic analysis. PhD Thesis, Santiago de Compostela University, Spain (in Spanish), 2003Google Scholar
  11. 11.
    Fernández E. A., Valtuille R., Presedo J., Willshaw P. Comparison of different methods for hemodialysis evaluations by means of ROC curves: from artificial intelligence to current methods. Clin. Nephrol. 64(3): 205–213, 2005PubMedGoogle Scholar
  12. 12.
    Fernández E. A., Valtuille R., Presedo J., Willshaw P. Comparison of standard and artificial neural network estimators of hemodialysis adequacy. Artif. Organs 29(2): 159–165, 2005PubMedCrossRefGoogle Scholar
  13. 13.
    Fernández E. A., Valtuille R., Willshaw P., Balzarini M. (2007) Molecular kinetics modeling in hemodialysis: on-line molecular monitoring and spectral analysis. ASAIO J. 53(5):582–586PubMedGoogle Scholar
  14. 14.
    Fernández E. A., Valtuille R., Willshaw P., Perazzo C. A. Using artificial intelligence to predict the equilibrated blood urea concentration. Blood Purif. 19(3): 271–285, 2001PubMedCrossRefGoogle Scholar
  15. 15.
    Ghu J., Yang J., Chen I. U., Lai Y. Prediction of equilibrated BUN by an artificial neural network in high efficient hemodialysis. Am. J. Kidney Dis. 3: 638–646, 1998Google Scholar
  16. 16.
    Gotch, F. A. Kinetic modeling in hemodialysis. In: Clinical Dialysis, 2nd edn., edited by A. R. Nissenson, R. N. Fine, and D. Gentile. Norwalk, CT: Appleton and Lange, 1990Google Scholar
  17. 17.
    Kaufman A. M., Schneditz D., Smye S., Polaschegg H. D., Levin N. W. Solute disequilibrium and multicompartmental modeling. Adv. Ren. Replace. Ther. 2(37): 319–329, 1995PubMedGoogle Scholar
  18. 18.
    Kennedy R. L., Yuchun L., Van Roy B., Reed C. D., Lippman R. P. Solving Data Mining Problems Through Pattern Recognition. New Jersey: Prentice Hall, 1998Google Scholar
  19. 19.
    Locatelli F. Dose of dialysis, convection and hemodialysis patients outcome – what the HEMO study doesn’t tell us: the European viewpoint. Nephrol. Dial. Transplant. 18:1061–1065, 2003PubMedCrossRefGoogle Scholar
  20. 20.
    Locatelli F., Hannedouche T., Jacobson S. et al. The effect of membrane permeability on ESRD: design of a prospective randomized multicentre trial. J. Nephrol. 12: 85–88, 1999PubMedCrossRefGoogle Scholar
  21. 21.
    Mevik B. H., Cederkvist H. R. Mean Squared Error of prediction (MSEP) estimates for principal component regression (PCR and partial least squares regression (PLSR). J. Chemomet. 18(9): 422–429, 2004CrossRefGoogle Scholar
  22. 22.
    NKF-K/DOQI Clinical Practice Guidelines for Hemodialysis Adequacy: update. Am. J. Kidney Dis. 48(S1):S2–90, 2006Google Scholar
  23. 23.
    Roa L. M., Prado M. The role of urea kinetic modeling in assessing the adequacy of dialysis. Crit. Rev. Biomed. Eng. 32(5–6): 461–539, 2004PubMedCrossRefGoogle Scholar
  24. 24.
    Section II. Haemodialysis adequacy. Nephrol. Dial. Transplant. 17(S7):16–31, 2002Google Scholar
  25. 25.
    Sharma A., Espinosa P., Bell L., Tom A., Rodd C. Multicompartmental urea kinetics in well-dialyzed children. Kidney Int. 58: 2138–2146, 2000PubMedCrossRefGoogle Scholar
  26. 26.
    Shawe-Taylor J., Kernel C. N. Methods for Pattern Analysis. Cambridge: Cambridge University Press, 2005Google Scholar
  27. 27.
    Smye S., Tattersal J., Will E. Modeling the post-dialysis rebound: the reconciliation of current formulas. ASAIO J. 45(6): 562–569, 1999PubMedCrossRefGoogle Scholar
  28. 28.
    Smye S. W., Will E. J., Lindley E. J. Postdialysis and Equilibrium urea concentrations. Blood Purif. 20: 189, 2002PubMedCrossRefGoogle Scholar
  29. 29.
    Tattersal J., Detakats D., Chamney P., Greenwood R., Farrington K. The post dialysis rebound: predicting and quantifying its effect on Kt/V. Kidney Int. 50(6): 2094–2102, 1996CrossRefGoogle Scholar
  30. 30.
    Wold S., Sjöström M., Eriksson L. PLS-regression: a basic tool of chemometrics. Chemomet. Intell. Lab. Syst. 58: 109–130, 2001CrossRefGoogle Scholar

Copyright information

© Biomedical Engineering Society 2008

Authors and Affiliations

  • E.A. Fernández
    • 1
    • 2
  • R. Valtuille
    • 3
  • P. Willshaw
    • 4
  • M. Balzarini
    • 2
    • 5
  1. 1.Faculty of EngineeringCatholic University of CórdobaCordobaArgentina
  2. 2.National Council of Scientific and Technological Research (CONICET)Buenos AiresArgentina
  3. 3.Fresenius Medical CareBuenos AiresArgentina
  4. 4.School of Health ScienceSwansea UniversitySwanseaUK
  5. 5.Biometric DepartmentNational University of CórdobaCordobaArgentina

Personalised recommendations